An auto-rotation module having a single-layer neural network on a user device can convert a document image to a monochrome image having black and white pixels and segment the monochrome image into bounding boxes, each bounding box defining a connected segment of black pixels in the monochrome image. The auto-rotation module can determine textual snippets from the bounding boxes and prepare them into input images for the single-layer neural network. The single-layer neural network is trained to process each input image, recognize a correct orientation, and output a set of results for each input image. Each result indicates a probability associated with a particular orientation. The auto-rotation module can examine the results, determine what degree of rotation is needed to achieve a correct orientation of the document image, and automatically rotate the document image by the degree of rotation needed to achieve the correct orientation of the document image.
Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
2. The method according to claim 1, wherein the converting comprises performing an adaptive binarization on the document image in which, for every pixel in the document image, a neighborhood of pixels is examined so as to separate characters in the document image from background of the document image or from one another.
3. The method according to claim 1, wherein the bounding boxes overlap one another and wherein the connected segments of pixels do not overlap with one another.
This invention relates to image processing techniques for analyzing and segmenting visual data, particularly in scenarios where overlapping bounding boxes and non-overlapping pixel segments are required. The method addresses the challenge of accurately identifying and isolating distinct regions within an image while ensuring that bounding boxes can overlap to capture multiple objects or features, but the connected segments of pixels within those regions do not overlap, preventing redundant or conflicting segmentation. The process involves detecting and defining bounding boxes around areas of interest in an image, where these boxes may overlap to account for adjacent or partially occluded objects. Within each bounding box, connected segments of pixels are identified and segmented such that no two segments overlap, ensuring clear and distinct separation of the image regions. This approach is useful in applications like object detection, medical imaging, or autonomous navigation, where precise segmentation is critical for accurate analysis. The method ensures that while bounding boxes can overlap to cover multiple objects or features, the pixel segments within those boxes remain non-overlapping, avoiding ambiguity in the segmentation results. This technique improves the accuracy and reliability of image analysis by maintaining distinct boundaries between segmented regions, even when the bounding boxes themselves intersect. The invention is particularly valuable in scenarios where overlapping regions are necessary but pixel-level precision is required.
7. The method according to claim 1, wherein the single-layer neural network is configured for outputting a probability value indicating that a zero degree of rotation or zero number of turns is needed to correct the orientation of the input textual snippet, a probability value indicating that a 90 degree of rotation or a single right or left turn is needed to correct the orientation of the input textual snippet, a probability value indicating that a 180 degree of rotation is or two right or left turns are needed to correct the orientation of the input textual snippet, and a probability value indicating that a 270 degree of rotation is or three right or left turns are needed to correct the orientation of the input textual snippet.
9. The system according to claim 8, wherein the converting comprises performing an adaptive binarization on the document image in which, for every pixel in the document image, a neighborhood of pixels is examined so as to separate characters in the document image from background of the document image or from one another.
10. The system according to claim 8, wherein the bounding boxes overlap one another and wherein the connected segments of pixels do not overlap with one another.
14. The system according to claim 8, wherein the single-layer neural network is configured for outputting a probability value indicating that a zero degree of rotation or zero number of turns is needed to correct the orientation of the input textual snippet, a probability value indicating that a 90 degree of rotation or a single right or left turn is needed to correct the orientation of the input textual snippet, a probability value indicating that a 180 degree of rotation is or two right or left turns are needed to correct the orientation of the input textual snippet, and a probability value indicating that a 270 degree of rotation is or three right or left turns are needed to correct the orientation of the input textual snippet.
This invention relates to a system for determining the optimal rotation or turning correction needed to properly orient textual snippets in digital images. The system addresses the problem of misaligned or incorrectly oriented text in scanned documents, photographs, or other digital images, which can hinder optical character recognition (OCR) accuracy and readability. The system uses a single-layer neural network to analyze the input textual snippet and outputs probability values for four possible correction actions: no rotation (0 degrees or zero turns), a 90-degree rotation or single right/left turn, a 180-degree rotation or two right/left turns, and a 270-degree rotation or three right/left turns. These probability values indicate the likelihood that each correction will properly orient the text. The neural network processes the input snippet to generate these probabilities, allowing the system to select the most appropriate correction based on the highest probability value. This approach improves text recognition and readability by automatically determining the correct orientation without manual intervention. The system is particularly useful in applications requiring high-accuracy text extraction from images, such as document digitization, automated data entry, and image-based text processing.
16. The computer program product according to claim 15, wherein the converting comprises performing an adaptive binarization on the document image in which, for every pixel in the document image, a neighborhood of pixels is examined so as to separate characters in the document image from background of the document image or from one another.
17. The computer program product according to claim 15, wherein the bounding boxes overlap one another and wherein the connected segments of pixels do not overlap with one another.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
June 14, 2021
November 22, 2022
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.